Blog

Measuring the DigiFarm Impact: Initial Insights from the Randomised Control Study

Can a mobile platform become a change agent for the world’s smallholder farmers by linking them to digital agricultural and financial services?

Read the Interim DigiFarm Impact Assessment report here.

AgriFin’s theory of change is that increased access to well-designed and bundled digital services drives gains for farmer income and productivity. As an innovation partner, AFA supported Safaricom through farmer-centric product development, business modeling, partnership support and impact measurement to build the DigiFarm platform.

DigiFarm is a free, integrated mobile platform that enables smallholder farmers and other agriculture value chain actors to access financial services, inputs, markets and other services. For smallholder farmers, DigiFarm provides access to three essential services: 

  1. Interactive information such as price and farming best practices
  2. Access to quality agricultural inputs at low-cost and convenient locations
  3. Access to small input loans linked directly to agricultural inputs supplied through the DigiFarm network. 

For agro-businesses, DigiFarm intends to offer near instant access to a large number of smallholder farmers at low cost, linking them to an efficient distribution system.

Building on the theme of being a change agent for smallholders and the environment around them, DigiFarm offers the opportunity for social impact and benefit for individual farmers, facilitating institutions, and the broader agricultural/farming ecosystem. Current rates of adoption indicate that DigiFarm could achieve the below reach and impact by 2023:

  •     4.35 million farmers will be subscribed to DigiFarm
  •     3.5 million farmers will be actively using the platform
  •     3.1 million farmers will be accessing learning content
  •     2.4 million farmers will be buying inputs
  •     900,000 farmers will be approved for loans to buy inputs.

The tools smallholders access through DigiFarm are expected to improve their income, productivity, and profitability, while building the capacity of commercial partners and institutions to provide financially sustainable smallholder services. To test this change hypothesis, Georgetown University Initiative on Innovation, Development and Evaluation (gui2de) has established randomized controlled trial (RCT) experiments in Nyandarua county to evaluate the impact of DigiFarm (i.e., the ‘MVP’, which provides a user with access to farm inputs, a learning platform, and Safaricom services) and DigiFarm loans (i.e., the ‘MVP+loans’, which offers a loan for input purchases in addition to the same suite of services as the MVP). 


 Figure 1: Diagram of the treatment arms for MVP RCT (left) and MVP+loans RCT (right)

The first experiment (Sample I) studies the MVP solution and randomly divides a sample of 3,152 households into four groups: MVP encouragement only, MVP and M-SHWARI encouragements, M-SHWARI encouragement only, and a pure control group that does not receive any product encouragement. This allows us to assess the impact that access to the basic package of DigiFarm services can have on farmers, in comparison to the impact that a simple stand-alone credit option (i.e., M-SHWARI, which is not tied to input purchases) might have, and to the joint effect of these two tools.

The second experiment (Study II) studies DigiFarm loans and randomly divides a sample of 3,003 households into two groups: MVP encouragement only, and MVP and DigiFarm loans encouragements. This second study allows us to measure the marginal impact of access to an integrated credit facility, in addition to any effect the MVP might produce. Whereas Sample I studies a loan that can be used for any purpose, Sample II looks at a loan that is specifically tied to the purchase of farm inputs.

For both of these experiments, gui2de hopes to evaluate the following outcomes of interest:

  • Registration
  • Productivity
  • Perceived quality of inputs
  • Income/revenue
  • Savings
  • Engagement in the learning platform
  • Agricultural knowledge
  • Usage (M-PESA, M-SHWARI, loan, etc.)
  • User experience 

Data collection started in February 2017 to establish a baseline before any product had been released. The following is a preliminary analysis based on data that had been collected two months after the intervention (i.e., product encouragement) period. The timing for the endline data collection period is still pending. 


Figure 2: Map of Sample I and Sample II households, Nyandarua County.

While the RCT is still ongoing, initial insights so far include the following:

Registration depends on level of education and to some extent gender

8 in 10 of those introduced to MVP registered for DigiFarm. The study has shown that those with no schooling and women are less likely to register. That is, 30% of farmers with no formal education and 21% of female farmers did not register.

Interacting on the learning platform is influenced by education and gender:

3 in 10 of those introduced to the MVP used the Arifu platform. Those having no formal education and females again are less likely to use the learning platform.

There is low adoption of DigiFarm loans (based on 2-3 months of data):

Less than 1 in 10 of households introduced to the MVP + loan applied for the DigiFarm loan, with data showing no differences in appeal across education levels and gender.

In conclusion

The DigiFarm platform has been undergoing enhancements, and the accessibility to iProcure shops is being improved. Due to very low rates of activity in the treatment group, even amongst registered users, no clear differences in behavior or outcomes could be detected at the time of this preliminary analysis. A final analysis at the end of the study will look at productivity, perceived quality of inputs, income, savings, and user experience. gui2de is also conducting A/B tests on a separate and much larger sample to nudge users to use either the learning platform or DigiFarm loan service. These tests will help us to further understand registration, engagement in the learning platform, and usage of DigiFarm loans.

Read the Interim DigiFarm Impact Assessment report here.

Comments are closed